Digital technology is anything but immaterial. Behind every ‘cloud’ or online service lies a vast chain of data centers spread across the globe, along with a dense network of infrastructure and end-user devices such as laptops and smartphones. The digital sector's carbon footprint is estimated to be on par with that of the aviation industry, accounting for 3% to 4% of global emissions, a figure consistently growing.
A sustainable IT approach ensures that IT services and infrastructures operate within the boundaries of both planetary and social limits. These constraints apply throughout the entire lifecycle of digital devices—from the extraction of raw materials needed for their production (as almost none of these materials come from recycling channels) through their usage and eventual disposal. For any research institution or funding organisation seeking to reduce its environmental footprint, the first crucial step is measuring it.
EPFL was among the first universities to estimate its digital footprint in 2019, thanks to its pioneering adoption of a sustainable IT approach. We established a dedicated team for sustainable IT and now include our digital footprint in our annual environmental report.
Digital technology is a textbook case of the challenges involved in measuring environmental footprints. Various calculators exist to assess the energy consumption of computing tasks, as Green Algorithm or Code Carbon, or AI services with Ecologits, among others. The digital sector has dedicated databases (here and there) and even its own certification label. While many tools are available, they do not necessarily make the task easier, so we must remain humble when interpreting environmental footprint estimates. As numerous assumptions are required to arrive at an overall measurement, the larger the institution, the less precise the inventory and consumption metrics become.
Nevertheless, we should not allow this to stop us: measurement is not an end in itself but a necessary tool for managing our environmental impact. The goal is not merely to obtain a precise final figure, in tons of CO2 equivalent, but to implement concrete actions to reduce it. Measuring our carbon footprint allows us to identify the most impactful categories and track progress in reducing it.
Estimating a consistent reduction in our footprint is particularly challenging. For example, in 2022, it would have been impossible to predict the rapid rise of generative AI for the general public. However, there are some levers already within our reach: sharing digital infrastructures between laboratories and universities, carefully selecting the models we use, optimising code for long-running simulations, adopting good and efficient data management practices etc. Many of these considerations are particularly important for universities of science and technology, operating at the forefront of digital infrastructures and where they are central to research and education. Working together through associations such as CESAER can support the exchange of lessons learnt and effective practices as these institutions navigate the best paths forward.
Digital infrastructures are nowadays at the heart of many areas of research and are also critical to the smooth operation of universities. In the face of ongoing and upcoming geopolitical crises – whether related to mining resources, equipment supply chains, or the availability of digital services – adopting a more frugal and judicious approach to the use of digital software and hardware infrastructure is crucial. Only by embracing digital sobriety can we sustain research excellence while navigating environmental, social, economic, and energy-related constraints. Reducing our usage, extending the lifespan of our IT infrastructures, and improving its efficiency remain essential.
Manuel Cubero-Castan
Project Manager in Sustainable IT at EPFL
We use both our own cookies and those of carefully selected partners we collaborate with.
Check out our detailed Cookie policy » and our Privacy policy » .